Related Experiment Videos
NARMAX representation and identification of ankle dynamics
Sunil L Kukreja1, Henrietta L Galiana, Robert E Kearney
1Division of Automatic Control, Department of Electrical Engineering, Linköpings Universitet, SE-581 83 Linköping, Sweden.
IEEE Transactions on Bio-Medical Engineering
|March 6, 2003
Summary
This study models ankle dynamics using a nonlinear autoregressive, moving average exogenous (NARMAX) approach. The NARMAX model accurately represents ankle behavior and provides precise parameter estimates from experimental data.
Area of Science:
- Biomechanics
- Systems Biology
- Computational Neuroscience
Background:
- Ankle dynamics are complex, involving nonlinear interactions.
- Existing models may not fully capture the intricacies of ankle joint behavior.
Purpose of the Study:
- To represent and identify ankle dynamics using a nonlinear autoregressive, moving average exogenous (NARMAX) model.
- To validate the NARMAX model's accuracy against continuous-time simulations and experimental data.
Main Methods:
- Derivation of a nonlinear difference equation for ankle dynamics.
- Application of NARMAX identification techniques to the derived model.
- Investigation of model properties using continuous-time simulations.
- Validation using experimental human ankle data.
Main Results:
- The NARMAX model closely matched outputs from continuous-time simulation methods.
- NARMAX identification yielded accurate discrete-time parameter estimates for ankle dynamics.
- High cross-validation variance accounted for in experimental human ankle data.
Conclusions:
- The NARMAX framework provides an effective method for modeling and identifying human ankle dynamics.
- The developed model demonstrates high accuracy and predictive capability for ankle joint behavior.
- This approach offers a robust tool for analyzing complex biomechanical systems.